Abstract :
Summary form only given, as follows. Development of Exascale Computers before the end of this decade is an ambitious target taken up by the international HPC community. The challenges are numerous; the important ones being power consumption & dissipation, reliability, application scaling, display technology and ubiquitous access. These challenges also occur at all levels of abstraction: materials & device physics, circuit design, logic design, architecture, the software stack, algorithms and high-speed networking (the last one for ubiquitous access). We need to meet these challenges not only at every level of abstraction but also need to look at the interaction between levels. Meeting the target of building Exascale systems will of course help us solve some "grand challenge" problems in science and engineering. More importantly, perhaps, it will help us "democratise" the use of HPC because of the affordability in terms of both capital and operating costs. However, such democratisation will require meeting yet another overriding challenge: education. A large workforce will be needed to design algorithms, optimise application software and intelligently use (high-end) simulation in their design flow. The talk will cover all the above aspects. Specific examples will be used for concrete illustration.
